COMMENT ⊗   VALID 00003 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	%mental[s90,jmc] Mental objects in the blocks world: designs
C00007 00003	\smallskip\centerline{Copyright \copyright\ 1990\ by John McCarthy}
C00008 ENDMK
C⊗;
%mental[s90,jmc] Mental objects in the blocks world: designs
\input memo.tex[let,jmc] \centerline{DRAFT DRAFT DRAFT}
\title{MENTAL OBJECTS IN THE BLOCKS WORLD: DESIGNS}


	%(re: mental situations, planning, design, towers)

	We treat an extension of the concept of the situation
calculus without actually changing the formalism.  The idea is
that situations have mental as well as physical components.  We
proceed as follows.

	There is a large supply of objects about which nothing
holds in the initial situation.  (In order to make this work, we
have to avoid axioms like % $$¬holds(p,s) ⊃ holds(not p,s).$$ %
Very likely we shall accomplish this by having more than one kind
of holds.)

	Suppose $x$ is one of these nondescript objects.  A
possible action is $declare(x,tower)$.  We have %
$$holds(is(x,tower),result(declare(x,tower),s)).$$

	The interpretation is that $x$ comes to be considered a
tower.  In the situation $result(declare(x,tower),s)$, $x$ is a
tower but has no other properties.  It has no location, it has no
blocks, and it has no physical existence.  Suppose $y$ is a
block.  There is an action $include(y,x)$.  Its effect is to
include the block $y$ in the tower $x$.  This action again has no
physical effect.  We have %
$$holds(in(y,x),result(include(y,x),s)).$$ % I haven't decided
about preconditions for such actions.  Our intent is to use it
when $x$ has been declared to be a tower and $y$ is a block.
However, I don't know whether we need to make this a precondition
for the action. $include(y,x)$ also doesn't locate $y$ at any
particular position.

	Our plan is to prescribe a set of actions, all mental,
adequate for designing a tower.  These actions need not obey
physical restrictions, and maybe it will be convenient to refrain
from building consistency conditions.  However, there will be a
fluent $proper.tower(x)$, and $holds(proper.tower(x),s)$ is
intended to mean that $x$ is a fully designed tower.  However, it
still needn't physically exist, because all the actions
building it were purely mental.

	Now we introduce physical actions.  They will be
essentially the conventional blocks world physical actions of
moving a block to a location or painting a block, and they will
have the usual physical preconditions.  However, these actions
may be prescribed to have effects relative to the abstract towers
that have been designed.  This will enable us to show that a
certain tower exists in a situation, i.e. $exists(x,s)$, or has
been partially constructed in a situation.  The latter may be
expressed by assertions of the existence of subtowers.

	Note: Leora Morgenstern pointed out that this treatment
of mental objects doesn't involve knowledge or belief.
\smallskip\centerline{Copyright \copyright\ 1990\ by John McCarthy}
\smallskip\noindent{\fiverm This draft of MENTAL[S90,JMC]\ TEXed on \jmcdate\ at \theTime}
%File originated on 06-Apr-90
\vfill\eject\end